RISC, CISC, and ISA Variations CS 3410 Computer System Organization & Programming These slides are the product of many rounds of teaching CS 3410 by Professors Weatherspoon, Bala, Bracy, and Sirer. Which is not considered part of the ISA? A. There is a control delay slot. B. The number of inputs each instruction can have. C. Load-use stalls will not be detected by the processor. D. The number of cycles it takes to execute a multiply. E. Each instruction is encoded in 32 bits. iClicker Question (first alone, then pairs) 2 Big Picture: Where are we now? Write BackMemoryFetch ExecuteDecode extend register file control alu memory din dout addr PC memory new pc in st IF/ID ID/EX EX/MEM MEM/WB im m B A ct rl ct rl ct rl B D D M compute jump/branch targets +4 forward unitdetect hazard 3 Big Picture: Where are we going? 4 int x = 10; x = x + 15; C compiler addi r5, r0, 10 addi r5, r5, 15 MIPS assembly 00100000000001010000000000001010 00100000101001010000000000001111 machine code assembler CPU Circuits Gates Transistors Silicon loader addi r0 r5 10 r0 = 0 r5 = r0 + 10 r5 = r15 + 15 A B 32 32RF addi r5, r0, 10 muli r5, r5, 2 addi r5, r5, 15 Big Picture: Where are we going? 5 int x = 10; x = 2 * x + 15; C compiler MIPS assembly machine code assembler CPU Circuits Gates Transistors Silicon loader 00100000000001010000000000001010 00000000000001010010100001000000 00100000101001010000000000001111 High Level Languages Instruction Set Architecture (ISA) Instruction Set Architectures • ISA Variations, and CISC vs RISC • Peek inside some other ISAs: • X86 • ARM Goals for Today 6 Instruction Set Architecture (ISA) Different CPU architectures specify different instructions Two classes of ISAs • Reduced Instruction Set Computers (RISC) IBM Power PC, Sun Sparc, MIPS, Alpha • Complex Instruction Set Computers (CISC) Intel x86, PDP-11, VAX • Another ISA classification: Load/Store Architecture • Data must be in registers to be operated on For example: array[x] = array[y] + array[z] 1 add ? OR 2 loads, an add, and a store ? • Keeps HW simple àmany RISC ISAs are load/store 7 iClicker Question What does it mean for an architecture to be called a load/store architecture? (A) Load and Store instructions are supported by the ISA. (B) Load and Store instructions can also perform arithmetic instructions on data in memory. (C) Loads & Stores are the primary means of reading and writing data in the ISA. (D) Data must first be loaded into a register before it can be operated on. (E) Every load must have an accompanying store at some later point in the program. 8 ISA defines the permissible instructions • MIPS: load/store, arithmetic, control flow, … • ARMv7: similar to MIPS, but more shift, memory, & conditional ops • ARMv8 (64-bit): even closer to MIPS, no conditional ops • VAX: arithmetic on memory or registers, strings, polynomial evaluation, stacks/queues, … • Cray: vector operations, … • x86: a little of everything ISA Variations 9 Accumulators • Early computers had one register! • Two registers short of a MIPS instruction! • Requires memory-based addressing mode - Example: add 200 // ACC = ACC + Mem[200] Brief Historical Perspective on ISAs 10 EDSAC (Electronic Delay Storage Automatic Calculator) in 1949 Intel 8008 in 1972 Next step: More Registers • Dedicated registers - separate accumulators for mult/div instructions • General-purpose registers - Registers can be used for any purpose -MIPS, ARM, x86 • Register-memory architectures - One operand may be in memory (e.g. accumulators) - x86 (i.e. 80386 processors) • Register-register architectures (aka load-store) - All operands must be in registers -MIPS, ARM Brief Historical Perspective on ISAs 11 • # of available registers plays huge role in ISA design ISAs are a product of current technology 12 Machine # General Purpose Registers Architectural Style Year EDSAC 1 Accumulator 1949 IBM 701 1 Accumulator 1953 CDC 6600 8 Load-Store 1963 IBM 360 18 Register-Memory 1964 DEC PDP-8 1 Accumulator 1965 DEC PDP-11 8 Register-Memory 1970 Intel 8008 1 Accumulator 1972 Motorola 6800 2 Accumulator 1974 DEC VAX 16 Register-Memory, Memory-Memory 1977 Intel 8086 1 Extended Accumulator 1978 Motorola 6800 16 Register-Memory 1980 Intel 80386 8 Register-Memory 1985 ARM 16 Load-Store 1985 MIPS 32 Load-Store 1985 HP PA-RISC 32 Load-Store 1986 SPARC 32 Load-Store 1987 PowerPC 32 Load-Store 1992 DEC Alpha 32 Load-Store 1992 HP/Intel IA-64 128 Load-Store 2001 AMD64 (EMT64) 16 Register-Memory 2003 People programmed in assembly and machine code! • Needed as many addressing modes as possible • Memory was (and still is) slow CPUs had relatively few registers • Register’s were more “expensive” than external mem • Large number of registers requires many bits to index Memories were small • Encouraged highly encoded microcodes as instructions • Variable length instructions, load/store, conditions, etc In the Beginning… 13 John Cock • IBM 801, 1980 (started in 1975) • Name 801 came from the bldg that housed the project • Idea: Can make a very small and very fast core • Known as “the father of RISC Architecture” • Turing Award and National Medal of Science Reduced Instruction Set Computer (RISC) 14 Dave Patterson • RISC Project, 1982 • UC Berkeley • RISC-I: ½ transistors & 3x faster • Influences: Sun SPARC, namesake of industry Reduced Instruction Set Computer (RISC) 15 John L. Hennessy • MIPS, 1981 • Stanford • Simple, full pipeline • Influences: MIPS computer system, PlayStation, Nintendo MIPS = Reduced Instruction Set Computer (RlSC) • ≈ 200 instructions, 32 bits each, 3 formats • all operands in registers - almost all are 32 bits each • ≈ 1 addressing mode: Mem[reg + imm] x86 = Complex Instruction Set Computer (ClSC) • > 1000 insns, 1-15 bytes each (dozens of add insns) • operands in dedicated registers, general purpose registers, memory, on stack, … - can be 1, 2, 4, 8 bytes, signed or unsigned • 10s of addressing modes - e.g. Mem[segment + reg + reg*scale + offset] RISC vs. CISC 16 RISC • Single-cycle execution • Hardwired control • Load/store architecture • Few memory addressing modes • Fixed-length insn format • Reliance on compiler optimizations • Many registers (compilers are better at using them) The RISC Tenets 17 CISC • many multicycle operations • microcoded multi-cycle operations • register-mem and mem-mem • many modes • many formats and lengths • hand assemble to get good performance • few registers RISC Philosophy Regularity & simplicity Leaner means faster Optimize common case Energy efficiency Embedded Systems Phones/Tablets RISC vs CISC 18 CISC Rebuttal Compilers can be smart Transistors are plentiful Legacy is important Code size counts Micro-code! “RISC Inside” Desktops/Servers What is one advantage of a CISC ISA? A. It naturally supports a faster clock. B. Instructions are easier to decode. C. The static footprint of the code will be smaller. D. The code is easier for a compiler to optimize. E. You have a lot of registers to use. iClicker Question (alone first, then pairs) 19 All MIPS instructions are 32 bits long, 3 formats R-type I-type J-type MIPS instruction formats 20 op rs rt rd shamt func 6 bits 5 bits 5 bits 5 bits 5 bits 6 bits op rs rt immediate 6 bits 5 bits 5 bits 16 bits op immediate (target address) 6 bits 26 bits All ARMv7 instructions are 32 bits long, 3 formats R-type I-type J-type ARMv7 instruction formats 21 opx op rs rd opx rt 4 bits 8 bits 4 bits 4 bits 8 bits 4 bits opx op rs rd immediate 4 bits 8 bits 4 bits 4 bits 12 bits opx op immediate (target address) 4 bits 4 bits 24 bits while(i != j) { if (i > j) i -= j; else j -= i; } Loop: BEQ Ri, Rj, End // if "NE" (not equal), stay in loop SLT Rd, Rj, Ri // (i > j) à Rd=1, (i ≤ j) à Rd = 0 BEQ Rd, R0, Else // Rd == 0 means (i ≤ j) à Else SUB Ri, Ri, Rj // i = i-j; J Loop Else: SUB Rj, Rj, Ri // j = j-i; J Loop End: MIPS Control Dependence 22 In MIPS, performance suffers if code has a lot of branches 3 NOP injections due to delay slot while(i != j) { if (i > j) i -= j; else j -= i; } Loop: CMP Ri, Rj // set condition registers // Example: 4, 3 à CR = 0101 // 5,5 à CR = 1000 SUBGT Ri, Ri, Rj // i = i-j only if CR & 0001 != 0 SUBLE Rj, Rj, Ri // j = j-i only if CR & 1010 != 0000 BNE loop // if "NE" (not equal), then loop ARMv7 Conditional Instructions 23 ARM: avoids delays with conditional instructions New: 1-bit condition registers (CR) = ≠ < > Control Independence! Shift one register (e.g., Rc) any amount Add to another register (e.g., Rb) Store result in a different register (e.g. Ra) ADD Ra, Rb, Rc LSL #4 Ra = Rb + Rc << 4 Ra = Rb + Rc x 16 ARMv7: Other Cool operations 24 ARMv7 instructions are 32 bits long, 3 formats Reduced Instruction Set Computer (RISC) properties • Only Load/Store instructions access memory • Instructions operate on operands in processor registers • 16 registers Complex Instruction Set Computer (CISC) properties • Autoincrement, autodecrement, PC-relative addressing • Conditional execution • Multiple words can be accessed from memory with a single instruction (SIMD: single instr multiple data) ARMv7 Instruction Set Architecture 25 ARMv8 instructions are 64 bits long, 3 formats Reduced Instruction Set Computer (RISC) properties • Only Load/Store instructions access memory • Instructions operate on operands in processor registers • 32 registers and r0 is always 0 Complex Instruction Set Computer (CISC) properties • Conditional execution • Multiple words can be accessed from memory with a single instruction (SIMD: single instr multiple data) ARMv8 (64-bit) Instruction Set Architecture 26 The number of available registers greatly influenced the instruction set architecture (ISA) Complex Instruction Set Computers were very complex + Small # of insns necessary to fit program into memory. - greatly increased the complexity of the ISA as well. Back in the day… CISC was necessary because everybody programmed in assembly and machine code! Today, CISC ISA’s are still dominant due to the prevalence of x86 ISA processors. However, RISC ISA’s today such as ARM have an ever increasing market share (of our everyday life!). ARM borrows a bit from both RISC and CISC. ISA Takeaways 27